Re: GNUstep Window Manager (was RE: Idea)

[Gregory Casamento]

--- "Philippe C.D. Robert" <phr@studiosendai.com> wrote:
> Hi,
> 
> I don't think this thread is of further use, at least not for this  
> list. So I will just comment one paragraph which I believe is a  
> common mistake. I hope you understand that.
> 
> > > Further, when I compare the results of GSBench (on a 400MHz PII  
> with
> > > an ultra fast graphics card) to those of NXBench (25MHz m68k) I get   
> > > just scared...;-)
> >
> > Are you sure this comparison is fair as it is ?  We have an additional 
> > layer, X, and X is very slow itself.
> 
> I cannot believe that X is so slow as many seem to believe (at least  
> not in those magnitudes). I could run X on a i486 machine with 16  
> Megs of RAM quite fine. What I cannot do is to run GS, GNOME, KDE or  
> whatever desktop on such a machine, thus I believe the problem is  
> mainly caused by such higher level technologies. Today we have  
> gigahertz machines, ultrafast graphics cards and more than enough  
> memory, I just refuse to believe that it is X and only X that makes  
> GNUstep so slow.
> The comparison is IMHO fair because I basically do not care about  
> 'under the hood issues' when I want to run an application: either it  
> is fast enough or not. The fact that GNUstep runs more or less as  
> fast on todays PCs as NEXTSTEP on old NeXTs (or slower in some  
> respects) shows me 2 things:
> 
> A) NEXTSTEP and the NeXT machines are brilliantly engineered gems.

I vehemently agree w/ this.

> B) GNUstep gui is slow, either caused by unoptimised implementations  
> and/or design flaws. I do not know the reason myself.
> 
> Thus - to go back to regular work - I believe we should care about  
> perfomance, otherwise we will never see applications like Quantrix,  
> Diagram or WetPaint on GNUstep - and we all agree that it's the  
> existence of such 'killer apps' that is required to make GNUstep more  
> popular.

I agree w/ this as well.   I believe that the reason for the gui library's
slowness is due to the fact that, up until now, we haven't been concerned about
performance.  Heavy usage/testing of the library is the only way to determine
where optimizations are needed.

> cheers, Phil
> 
> --
> Philippe C.D. Robert | Senior Developer
> http://www.groupville.com - GroupWare made easy!
> 
> _______________________________________________
> Discuss-gnustep mailing list
> Discuss-gnustep@gnu.org
> http://mail.gnu.org/mailman/listinfo/discuss-gnustep

Later, 
Greg "Going back to work on the NSSpellServer" Casamento

=====
======================================================
Gregory John Casamento           Sr. Software Engineer
======================================================
Software patents are endangering the computer industry
in the US and around the world.  Join the LPF and help 
the cause.       http://lpf.ai.mit.edu/
======================================================

__________________________________________________
Do You Yahoo!?
Yahoo! Photos - Share your holiday photos online!
http://photos.yahoo.com/